Myoelectric partial-hand prostheses offer the potential for a wide range of functional hand grasps not previously available to partial-hand amputees. Despite these advances, there is still an important need to develop more effective methods of controlling them. Many interesting studies investigating the use of pattern- recognition-based myoelectric control (PRMC) have shown that electromyographic signals (EMG) from the extrinsic muscles of the hand provide highly-accurate control of prostheses in individuals with high-level amputations. However, these studies fail to address an essential challenge that is unique to partial-hand amputees: the presence of a functional wrist. Not only can the wrist be in any position when the user initiates movement of their prosthesis, it can also be in active motion during the movements. Moreover, the muscle contractions responsible for these motions influence properties of the EMG from the extrinsic muscles, and consequently the performance of a PRMC system. In this study, we will use principal component analysis to describe how wrist kinematics affect the properties of EMG from the extrinsic muscles. We will incorporate recorded wrist kinematic information into both linear and non-linear pattern-recognition systems. Finally, we will test the real-time ability of these PRMC systems to accommodate the effect of wrist kinematics, preserve wrist function and improve partial-hand prosthetic control.